Load typed data

Replace the placeholders with values from your own authorized scope before copying this prompt into an MCP client.

Inputs

• {{workspace_id}} Stable workspace ID.
• {{application_id}} Stable application ID.
• {{scope_members}} Nonempty map restricting the exact load scope.
• {{typed_rows}} Full-coordinate typed rows, with decimals supplied as strings.
• {{request_key}} Stable source-batch begin key.

Prompt

For {{workspace_id}} / {{application_id}}, use read_v2_program with application.state.read to get metadataRevision and dataRevision.
Use execute_v2_source_program with facts.source.begin, mode merge, scopeMembers {{scope_members}}, scopeAll false, expectedMetadataRevision, expectedDataRevision and idempotencyKey {{request_key}}.
Append {{typed_rows}} using facts.source.append with the returned batchId, chunkIndex starting at 0, and 1 to 500 rows per chunk. Every row has members, a typed value, and sourceRef.
Seal with facts.source.seal using the final acknowledged rowCount as expectedRows and contentDigest as expectedContentDigest. Run facts.source.preview with the same batchId and allowWarnings false.
Return the batchId, rowCount, contentDigest, setCount, clearCount and any diagnostics returned. Stop at review before facts.source.publish or any transfer execution request.
Stop earlier if the rows, scope, digest or revisions disagree.

Example inputs

For the disposable Operating Data Cube, use the row's members as scope_members. The preview returns one set and zero clears. If the cell starts at 10.00000000, a read still returns that value after preview.

[
  {
    "members": {
      "Account": "InputA", "Entity": "PointA", "Year": "Y2027",
      "Period": "P01", "View": "Input", "Value": "Raw", "Custom01": "ItemA"
    , "Scenario":"Actual", "DataType":"InputData"},
    "value": {"type": "AmountLiteral", "decimal": "12.50000000"},
    "sourceRef": "documentation-typed-load"
  }
]

Tools used include read_v2_program, execute_v2_source_program. See the module guidelines and tool index.